Image Formats
EVAnalyzer reads images through scopeio, a pure-Rust reader for microscopy and whole-slide images built into the application - no Java needed. Samples of 8, 16 and 32 bits (unsigned or signed) and 32/64-bit floating point are read, greyscale or RGB. When a file declares fewer significant bits than its sample type holds (e.g. 12-bit data stored as 16-bit), the value range follows the significant bits.
Supported Formats
Section titled “Supported Formats”| Format | Extension(s) |
|---|---|
| TIFF / BigTIFF / OME-TIFF | .tif .tiff .tf2 .tf8 .btf |
| Zeiss CZI / LSM / ZVI | .czi .lsm .zvi |
| Nikon ND2 | .nd2 |
| Leica LIF / LOF / XLEF / SCN | .lif .lof .xlef .xlcf .xlif .scn |
| Olympus VSI / OIB / OIF / OIR | .vsi .oib .oif .oir |
| Aperio SVS | .svs |
| Hamamatsu NDPI | .ndpi .ndpis |
| 3DHistech MIRAX | .mrxs |
| Ventana BIF, Akoya QPTIFF | .bif .qptiff |
| Imaris, DeltaVision, MetaMorph | .ims .dv .r3d .stk .nd |
| DICOM | .dcm .dicom |
| PNG, JPEG, JPEG 2000, BMP | .png .jpg .jpeg .jp2 .j2k .bmp |
Only files with these extensions are listed in the Images panel. Companion files that belong to a main file - such as the .ets files next to a .vsi - are opened through the main file.
Image Planes
Section titled “Image Planes”A single microscopy file typically contains many individual images organized into three dimensions:
| Dimension | Symbol | Description |
|---|---|---|
| Channel | C | Fluorescence or brightfield channel |
| Z-plane | Z | Focal plane in a z-stack |
| Time frame | T | Frame in a time-lapse sequence |
Each unique combination of (C, Z, T) is one image plane. EVAnalyzer can access and process any plane individually.
Pipeline input is specified by channel index (0-based). Z and T handling is configured in the Image control tab on the right hand side.
Metadata
Section titled “Metadata”EVAnalyzer takes from the file, per series:
- the number of channels, Z planes and time frames,
- each channel’s name and emission wavelength (its color in the viewer) - a channel without a name is shown as Channel 0, Channel 1, …,
- the physical pixel size in X, Y and Z,
- the frame interval of a time-lapse - the file’s time increment, or else the mean step between the time stamps of the first and the last frame,
- the objective’s magnification.
A wrong value can be corrected in the image info. RGB images are split into Red, Green and Blue channels (alpha is dropped).
Big Images
Section titled “Big Images”Large whole-slide images are handled automatically via tiling. EVAnalyzer splits any image larger than the configured tile size into overlapping tiles, analyses them independently, and stitches the results.
The tile size used for analysis is 4096 × 4096 px.
Cross-tile object merging
Section titled “Cross-tile object merging”An object that straddles an internal tile boundary (for example a whole organ or tissue region on a whole-slide image) would otherwise be exported as two or more separate fragments, one per tile. EVAnalyzer reassembles these back into a single correct object once every tile of an image has finished processing, recomputing its geometry (area, perimeter, ellipse fit, …) from the true merged shape rather than from either fragment alone - regardless of how many tiles the object spans.
This is on by default - most users have no reason to know or care about tiles, and just expect objects to be detected correctly regardless of where a tile boundary happens to fall.
See Cross-Tile Object Merging for how this works, why fixed-overlap tile stitching can’t solve the same problem, and the full settings reference.
Large images that come with pyramid levels (for example SVS, NDPI, MRXS, CZI, OME-TIFF) are read at the level that fits the zoom; see Viewing large images.